From my understanding, there arent many one-offs, but these updates. There are so many minor things I see in the release notes, that would stop some department from functioning. AP, AR, MRP, Order…so what are you doing on patches…how can you test all the functions before rolling out a patch? Do you just roll them out and let the users report?
When I’m helping a company with implementation, we build a set of test scripts to run transactions through each process flow to make sure the configuration works the way they want. Later, that same set of test scripts can be used again and again to test patches and upgrades. This catches most of the big issues and avoids disrupting business as much as possible.
Building the test scripts now will take some effort; however, you’ll have them to use in the future too, and you get the benefits of not annoying the users and customers.
Scripts using Epicor’s ATE or another tool?
It really depends on how extensive the scripts are. Epicor’s ATE can make it much easier to run them quickly. However, most of my clients do the test scripts manually, in part to have the key users involved in the testing (more buy-in). The intent is to catch any big breaks in processes or errors before deployment.
Thanks Jeff. I am VERY interested in automated testing but I am assuming you’re talking about User Testing with written scripts then?
I’d like to not only test code changes but making sure that users are following the implementation guidelines and identifying changes in process as well.